Advertisement

Ultrasound Image–Based Deep Features and Radiomics for the Discrimination of Small Fat-Poor Angiomyolipoma and Small Renal Cell Carcinoma

  • Author Footnotes
    1 These authors contributed equeally to this work and share co-first author.
    Li Zhang
    Footnotes
    1 These authors contributed equeally to this work and share co-first author.
    Affiliations
    Department of Ultrasound, Peking University Third Hospital, Beijing, China
    Search for articles by this author
  • Author Footnotes
    1 These authors contributed equeally to this work and share co-first author.
    Kui Sun
    Footnotes
    1 These authors contributed equeally to this work and share co-first author.
    Affiliations
    Department of Radiology, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Jinan, Shandong, China
    Search for articles by this author
  • Liting Shi
    Affiliations
    Department of Radiology, Shandong First Medical University and Shandong Academy of Medical Sciences, Taian, China
    Search for articles by this author
  • Jianfeng Qiu
    Affiliations
    Medical Science and Technology Innovation Center, Shandong First Medical University and Shandong Academy of Medical Sciences, Jinan, China
    Search for articles by this author
  • Ximing Wang
    Affiliations
    Department of Radiology, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Jinan, Shandong, China
    Search for articles by this author
  • Shumin Wang
    Correspondence
    Address correspondence to: Shumin Wang, Department of Ultrasound, Peking University Third Hospital, 49 North Garden Road, Haidian District, Beijing 100191, China.
    Affiliations
    Department of Ultrasound, Peking University Third Hospital, Beijing, China
    Search for articles by this author
  • Author Footnotes
    1 These authors contributed equeally to this work and share co-first author.

      Abstract

      We evaluated the performance of ultrasound image–based deep features and radiomics for differentiating small fat-poor angiomyolipoma (sfp-AML) from small renal cell carcinoma (SRCC). This retrospective study included 194 patients with pathologically proven small renal masses (diameter ≤4 cm; 67 in the sfp-AML group and 127 in the SRCC group). We obtained 206 and 364 images from the sfp-AML and SRCC groups with experienced radiologist identification, respectively. We extracted 4024 deep features from the autoencoder neural network and 1497 radiomics features from the Pyradiomics toolbox; the latter included first-order, shape, high-order, Laplacian of Gaussian and Wavelet features. All subjects were allocated to the training and testing sets with a ratio of 3:1 using stratified sampling. The least absolute shrinkage and selection operator (LASSO) regression model was applied to select the most diagnostic features. Support vector machine (SVM) was adopted as the discriminative classifier. An optimal feature subset including 45 deep and 7 radiomics features was screened by the LASSO model. The SVM classifier achieved good performance in discriminating between sfp-AMLs and SRCCs, with areas under the curve (AUCs) of 0.96 and 0.85 in the training and testing sets, respectively. The classifier built using deep and radiomics features can accurately differentiate sfp-AMLs from SRCCs on ultrasound imaging.

      Key Words

      To read this article in full you will need to make a payment

      Purchase one-time access:

      Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online access
      One-time access price info
      • For academic or personal research use, select 'Academic and Personal'
      • For corporate R&D use, select 'Corporate R&D Professionals'

      Subscribe:

      Subscribe to Ultrasound in Medicine and Biology
      Already a print subscriber? Claim online access
      Already an online subscriber? Sign in
      Institutional Access: Sign in to ScienceDirect

      References

        • Acharya AS
        • Prakash A
        • Saxena P
        • Nigam A.
        Sampling: Why and how of it?.
        Indian J Med Specialities. 2013; 4: 330-333
        • Aerts HJ
        • Velazquez ER
        • Leijenaar RT
        • Parmar C
        • Grossmann P
        • Carvalho S
        • Bussink J
        • Monshouwer R
        • Haibe-Kains B
        • Rietveld D
        • Hoebers F
        • Rietbergen MM
        • Leemans CR
        • Dekker A
        • Quackenbush J
        • Gillies RJ
        • Lambin P.
        Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach.
        Nat Commun. 2014; 5: 4006
        • Ball MW
        • Bezerra SM
        • Gorin MA
        • Cowan M
        • Pavlovich CP
        • Pierorazio PM
        • Netto GJ
        • Allaf ME.
        Grade heterogeneity in small renal masses: Potential implications for renal mass biopsy.
        J Urol. 2015; 193: 36-40
        • Cao H
        • Fang L
        • Chen L
        • Zhan J
        • Diao X
        • Liu Y
        • Lu C
        • Zhang Z
        • Chen Y.
        The independent indicators for differentiating renal cell carcinoma from renal angiomyolipoma by contrast-enhanced ultrasound.
        BMC Med Imaging. 2020; 20: 32
        • Conti A
        • Duggento A
        • Indovina I
        • Guerrisi M
        • Toschi N.
        Radiomics in breast cancer classification and prediction.
        Semin Cancer Biol. 2021; 72: 238-250
        • Dhungel N
        • Carneiro G
        • Bradley A.
        The automated learning of deep features for breast mass classification from mammograms.
        Lecture Notes Comput Sci. 2016; 9901 (In: Ourselin S, Joskowicz L, Sabuncu M, Unal G, Wells W, eds. Medical Image Computing and Computer-Assisted Intervention—MICCAI 2016): 106-114
        • Feng Z
        • Rong P
        • Cao P
        • Zhou Q
        • Zhu W
        • Yan Z
        • Liu Q
        • Wang W.
        Machine learning-based quantitative texture analysis of CT images of small renal masses: Differentiation of angiomyolipoma without visible fat from renal cell carcinoma.
        Eur Radiol. 2018; 28: 1625-1633
        • Fujii Y
        • Komai Y
        • Saito K
        • Iimura Y
        • Yonese J
        • Kawakami S
        • Ishikawa Y
        • Kumagai J
        • Kihara K
        • Fukui I.
        Incidence of benign pathologic lesions at partial nephrectomy for presumed RCC renal masses: Japanese dual-center experience with 176 consecutive patients.
        Urology. 2008; 72: 598-602
        • Hindman N
        • Long N
        • Genega EM
        • Melamed J
        • Pedrosa I.
        Angiomyolipoma with minimal fat: Can It Be Differentiated from clear cell renal cell carcinoma by using standard MR techniques?.
        Radiology. 2012; 265: 468-477
        • Hodgdon T
        • Mcinnes M
        • Schieda N
        • Flood TA
        • Lamb L
        • Thornhill RE.
        Can quantitative CT texture analysis be used to differentiate fat-poor renal angiomyolipoma from renal cell carcinoma on unenhanced CT images?.
        Radiology. 2015; 276: 787-796
        • Israel GM
        • Hindman N
        • Hecht E
        • Krinsky G.
        The use of opposed-phase chemical shift MRI in the diagnosis of renal angiomyolipomas.
        Am J Roentgenol. 2005; 184: 1868-1872
        • Jinzaki M
        • Tanimoto A
        • Narimatsu Y
        • Ohkuma K
        • Murai M.
        Angiomyolipoma: Imaging findings in lesions with minimal fat.
        Radiology. 1997; 205: 497-502
        • Jinzaki M
        • Silverman SG
        • Akita H
        • Nagashima Y
        • Mikami S
        • Oya M.
        Renal angiomyolipoma: A radiological classification and update on recent developments in diagnosis and management.
        Abdom Imaging. 2014; 39: 588-604
        • Katabathina VS
        • Vikram R
        • Nagar AM
        • Tamboli P
        • Menias CO
        • Prasad SR.
        Mesenchymal neoplasms of the kidney in adults: Imaging spectrum with radiologic–pathologic correlation.
        Radiographics. 2010; 30: 1525-1540
        • Kutikov A
        • Fossett LK
        • Ramchandani P
        • Tomaszewski JE
        • Siegelman ES
        • Banner MP
        • Arsdalen K
        • Wein AJ
        • Malkowicz SB.
        Incidence of benign pathologic findings at partial nephrectomy for solitary renal mass presumed to be renal cell carcinoma on preoperative imaging.
        Urology. 2006; 68: 737-740
        • Lai HY
        • Chen CK
        • Lee YH
        • Tsai PP
        • Chen JH
        • Shen WC.
        Multicentric aggressive angiomyolipomas: A rare form of PEComas.
        AJR Am J Roentgenol. 2006; 186: 837-840
        • Le MH
        • Chen J
        • Wang L
        • Wang Z
        • Liu W
        • Cheng KT
        • Yang X.
        Automated diagnosis of prostate cancer in multi-parametric MRI based on multimodal convolutional neural networks.
        Phys Med Biol. 2017; 62: 6497-6514
        • Lee HS
        • Hong H
        • Jung DC
        • Park S
        • Kim J.
        Differentiation of fat-poor angiomyolipoma from clear cell renal cell carcinoma in contrast-enhanced MDCT images using quantitative feature classification.
        Med Phys. 2017; 44: 3604-3614
        • Lee H
        • Hong H
        • Kim J
        • Jung DC.
        Deep feature classification of angiomyolipoma without visible fat and renal cell carcinoma in abdominal contrast-enhanced CT images with texture image patches and hand-crafted feature concatenation.
        Med Phys. 2018; 45: 1550-1561
        • Lovric M.
        International encyclopedia of statistical science.
        Springer, Berlin/Heidelberg2011
        • Lu Q
        • Wang W
        • Huang B
        • Li C
        • Li C.
        Minimal fat renal angiomyolipoma: The initial study with contrast-enhanced ultrasonography.
        Ultrasound Med Biol. 2012; 38: 1896-1901
        • Melzer A
        • Immel E
        • Michitsch S
        • Konak S
        • Bremer P.
        Sonographische Differenzierungskriterien von kleinen Angiomyolipomen (≤3 cm) und Nierenzellkarzinomen.
        RöFo. 1998; 169: 627-632
        • Moch H
        • Cubilla AL
        • Humphrey PA
        • Reuter VE
        • Ulbright TM.
        The 2016 WHO Classification of Tumours of the Urinary System and Male Genital Organs—Part A: Renal, penile, and testicular tumours.
        Eur Urol. 2016; 70: 93-105
        • Park BK.
        Renal angiomyolipoma: Radiologic classification and imaging features according to the amount of fat.
        AJR Am J Roentgenol. 2017; 209: 1-10
        • Park KJ
        • Kim MH
        • Kim JK
        • Cho KS.
        Sonographic features of small (<4 cm) renal tumors with low signal intensity on T2-weighted MR images: Differentiating minimal-fat angiomyolipoma from renal cell carcinoma.
        AJR Am J Roentgenol. 2018; 211: 605-613
        • Remzi M
        • Ozsoy M
        • Klingler HC
        • Susani M
        • Waldert M
        • Seitz C
        • Schmidbauer J
        • Marberger M.
        Are small renal tumors harmless? Analysis of histopathological features according to tumors 4 cm or less in diameter.
        J Urol. 2006; 176: 896-899
        • Richmond L
        • Atri M
        • Sherman C
        • Sharir S.
        Renal cell carcinoma containing macroscopic fat on CT mimics an angiomyolipoma due to bone metaplasia without macroscopic calcification.
        Br J Radiol. 2014; 83: 179-181
        • Sagreiya H
        • Akhbardeh A
        • Li D
        • et al.
        Point shear wave elastography using machine learning to differentiate renal cell carcinoma and angiomyolipoma[J].
        Ultrasound in Medicine & Biology. 2018; 45
        • Sasiwimonphan K
        • Takahashi N
        • Leibovich BC
        • Carter RE
        • Atwell TD
        • Kawashima A.
        Small (<4 cm) renal mass: Differentiation of angiomyolipoma without visible fat from renal cell carcinoma utilizing MR imaging.
        Radiology. 2012; 263: 160-168
        • Schwarze V
        • Froelich MF
        • Marschner C
        • Knsel T
        • Rübenthaler J
        • Clevert DA.
        Safe and pivotal approaches using contrast-enhanced ultrasound for the diagnostic workup of non-obstetric conditions during pregnancy, a single-center experience.
        Arch Gynecol Obstet. 2021; 303: 103-112
        • Shen W
        • Mu Z
        • Yang F
        • Yang C
        • Jie T.
        Multi-scale convolutional neural networks for lung nodule classification.
        Inf Process Med Imaging. 2015; 24: 588-599
        • Sherman J
        • Hartman D
        • Friedman A
        • Madewell J
        • Davis C
        • Angiomyolipoma Goldman S.
        Computed tomographic–pathologic correlation of 17 cases.
        AJR Am J Roentgenol. 1981; 137: 1221-1226
        • Siegel CL
        • Middleton WD
        • Teefey SA
        • Mcclennan BL.
        Angiomyolipoma and renal cell carcinoma: US differentiation.
        Radiology. 1996; 198: 789-793
        • Sim JS
        • Seo CS
        • Kim SH
        • Kim JH
        • Choi BI.
        Differentiation of small hyperechoic renal cell carcinoma from angiomyolipoma: Computer-aided tissue echo quantification.
        J Ultrasound Med. 1999; 18: 261-264
        • Sun R
        • Limkin EJ
        • Vakalopoulou M
        • Dercle L
        • Champiat S
        • Han SR
        • Verlingue L
        • Brandao D
        • Lancia A
        • Ammari S
        • Hollebecque A
        • Scoazec JY
        • Marabelle A
        • Massard C
        • Soria JC
        • Robert C
        • Paragios N
        • Deutsch E
        • Ferté C.
        A radiomics approach to assess tumour-infiltrating CD8 cells and response to anti-PD-1 or anti-PD-L1 immunotherapy: An imaging biomarker, retrospective multicohort study.
        Lancet Oncol. 2018; 19: 1180-1191
        • van Griethuysen JJM
        • Fedorov A
        • Parmar C
        • Hosny A
        • Aucoin N
        • Narayan V
        • Beets-Tan RGH
        • Fillion-Robin JC
        • Pieper S
        • Aerts H.
        Computational radiomics system to decode the radiographic phenotype.
        Cancer Res. 2017; 77: e104-e107
        • Wang C
        • Elazab A
        • Wu J
        • Hu Q.
        Lung nodule classification using deep feature fusion in chest radiography.
        Comput Med Imaging Graph. 2017; 57: 10-18
        • Yan L
        • Liu Z
        • Wang G
        • Huang Y
        • Liu Y
        • Yu Y
        • Liang C.
        Angiomyolipoma with minimal fat: Differentiation from clear cell renal cell carcinoma and papillary renal cell carcinoma by texture analysis on CT images.
        Acad Radiol. 2015; 22: 1115-1121
        • Yang R
        • Wu J
        • Sun L
        • Lai S
        • Xu Y
        • Liu X
        • Ma Y
        • Zhen X.
        Radiomics of small renal masses on multiphasic CT: Accuracy of machine learning-based classification models for the differentiation of renal cell carcinoma and angiomyolipoma without visible fat.
        Eur Radiol. 2020; 30: 1254-1263
        • Zabihollahy F
        • Schieda N
        • Krishna S
        • Ukwatta E.
        Automated classification of solid renal masses on contrast-enhanced computed tomography images using convolutional neural network with decision fusion.
        Eur Radiol. 2020; 30: 5183-5190
        • Zabor EC
        • Furberg H
        • Lee B
        • Campbell S
        • Lane BR
        • Thompson RH
        • Antonio EC
        • Noyes SL
        • Zaid H
        • Jaimes EA
        • Russo P.
        Long-term renal function recovery following radical nephrectomy for kidney cancer: Results from a multicenter confirmatory study.
        J Urol. 2018; 199: 921-926
        • Zhang YY
        • Luo S
        • Liu Y
        • Xu RT.
        Angiomyolipoma with minimal fat: Differentiation from papillary renal cell carcinoma by helical CT.
        Clin Radiol. 2013; 68: 365-370